Struct Ucb1 

pub struct Ucb1 { /* private fields */ }

A deterministic UCB1 multi-armed bandit.

Balances exploration (try arms whose value is uncertain) and exploitation (favour the arm with the best observed reward) using the classic upper confidence bound x̄ᵢ + c·√(ln N / nᵢ). Unpulled arms are tried first. No randomness, so the same reward stream always yields the same decisions — which is what makes the adaptive layer reproducible.

Implementations

impl Ucb1

pub const DEFAULT_EXPLORATION: f64 = std::f64::consts::SQRT_2

Default exploration constant c = √2, the textbook UCB1 value.

pub fn new(num_arms: usize) -> Self

Creates a bandit with num_arms arms (clamped to at least 1) and the default exploration constant.

pub fn with_exploration(num_arms: usize, exploration: f64) -> Self

Creates a bandit with an explicit exploration constant c (higher = more exploration).

pub fn arm_count(&self) -> usize

The number of arms.

pub fn pulls(&self, arm: usize) -> u64

Number of times arm has been pulled (0 if out of range).

pub fn select(&self) -> usize

Selects the next arm to try. Any never-pulled arm is chosen first (forced initial exploration); otherwise the arm with the highest UCB score.

pub fn record(&mut self, arm: usize, reward: f64)

Records a reward for arm (incremental mean update). Out-of-range arms and non-finite rewards are ignored.

pub fn best_arm(&self) -> Option<usize>

The current best arm by observed mean reward (pure exploitation), or None until at least one arm has been pulled.

pub fn mean_reward(&self, arm: usize) -> Option<f64>

The observed mean reward for arm, or None if it hasn’t been pulled.

Trait Implementations

impl Clone for Ucb1

fn clone(&self) -> Ucb1

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for Ucb1

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.